Evaluation of a sparse sampling strategy for determining vancomycin pharmacokinetics in preterm neonates: application of optimal sampling theory

OBJECTIVE: To use optimal sampling theory to determine the fewest vancomycin concentrations required and the appropriate sampling times to calculate vancomycin pharmacokinetic parameters in neonates. DESIGN: Unblinded evaluation in neonates with presumed sepsis. SETTING: Level 3 community-based neonatal intensive care unit. PATIENTS: Eleven neonates with presumed sepsis. INTERVENTIONS: Twelve courses of intravenous vancomycin 20 mg/kg were administered. Blood samples were collected 3 and 9 hours after initiation of a 1-hour infusion of the first dose. MEASUREMENTS AND MAIN RESULTS: A two-compartment model was fit to vancomycin concentrations using iterative two-stage analysis. Pharmacokinetic parameter estimates were used for determination of optimal sampling times for two-, three-, and four-sample strategies with subsequent generation of two-, three-, and four-sample concentration data for 100 cases. Relative performance of strategies was compared through calculation and comparison of D efficiency for the determined strategies. Bias (median percent error) and precision (median percent absolute error) of pharmacokinetic parameter estimates for each strategy in the 100 simulated cases were determined. CONCLUSIONS: For estimation of total clearance and volume in the central and peripheral compartments, all strategies performed similarly with no difference in efficiency or bias and precision of estimates. Our results suggest that for clinical evaluations two appropriately timed samples (0.5 h after a 1-h infusion, trough concentration) are adequate for estimation of vancomycin clearance in neonates.